Stimulation of the lumbar spinal cord allows rats and now people with paralysis to move their legs again. Such reanimation with spinal epidural stimulation can also be applied for recovery of hand function. As we seek to determine the mechanisms of this therapy, we are limited by the technology to apply stimulation to the cervical spinal cord of the rat, an important model for CNS injury and repair. The cervical spinal cord rotates and bends with head movement, so arrays must be supple. However, the electrode array also needs to be stiff, in order to place it into the thin epidural space. Current arrays, designed for the lumbar spinal cord, are made of Parylene-C, which is relatively stiff, making them likely to injure the cervical cord or to lose contact with the underlying dura mater. We have developed an array made of a softening polymer that is stiff at room temperature when dry and then becomes as supple as silicone rubber when implanted into the wet and warm body environment. In addition to softening at high temperatures, the arrays have electrodes patterned with photolithography, which allows high precision. We hypothesize that softening spinal stimulators will be safer and more effective at exciting spinal circuits than Parylene-C arrays. To test the safety of the arrays we will train rats on a food manipulation task that is sensitive for forepaw impairment. We will implant half of rats with softening polymer arrays and half with Parylene-C arrays. We will measure forepaw dexterity after implantation. Rats will be perfused, and the spinal cords examined for histological markers of inflammation and injury. Preliminary studies show that after the arrays are implanted in the epidural space dorsal to the cervical spinal cord, rats have no impairment in paw function. The arrays take the shape of the underlying spinal cord, and tissue markers of inflammation indicate no damage to the spinal cord. To test the efficacy of the arrays, we will measure electrode impedance and the stimulation intensity necessary to cause a muscle response-the spinal threshold-over 8 weeks of daily stimulation. We will also test the ability of implanted arrays to modulate muscle responses to motor cortex stimulation, as we demonstrate in our preliminary results. Finally, we will test the ability of paired brain and spinal cord stimulation to promote recovery of motor function after injury to the corticospinal tract. We expect that the softening polymer arrays will maintain a tight neural interface and will resist damage because they are flexible. This will resul in more effective long-term activation of spinal circuits and better functional recovery than Parylene-C arrays. Thus, we intend to fill a gap in technology for stimulating the cervical cord of the awake, behaving rat. This tool could dramatically accelerate our understanding of an important therapy to restore movement in people with paralysis.